The problem
It can be difficult for the average user of a social media platform to identify if a post on the climate crisis is based on a reliable source or misinformation. As a consequence, erroneous ‘news’ are reproduced or go viral. Some users even intentionally mislead the internet community. This behaviour can cause confusion in raising awareness and taking action on climate crisis.
The solution
@IsThisAnExpert is a Twitter bot that helps users see the credibility of another user on the climate crisis. Retweet a post on climate change to @IsThisAnExpert and it will run an assessment on the credibility of the user and their expertise on climate change.
A user’s credibility score is calculated by tweets and retweets of credible sources. A scientist affiliated with the IPCC, a reputable institution, many publications and citations (therefore a h-index) will have a high credibility score. A random user that has frequently retweeted the IPCC report will be assigned with a higher credibility score than Donald Trump who does not make references to credible sources. @IsThisAnExpert-Bots can be developed for a variety of social media platforms and even for other topics to raise awareness of the credibility of users. As credibility score is a retweet, it is publicly visible to other users in the thread and can be shared easily. It is likely to trigger interest and is easy to use. We therefore think that its outreach potential is large.
Accomplishments
The team met during the TheClimateChallenge hackathon and developed the idea of @IsThisAnExpert. We have created a formula for the credibility score, collated a list of climate change experts on Twitter users, computed these into a mySQL-based database and backend and merged with the twitter bot. Our bot is able to assign a credibility score for any twitter user and reply with a custom message that makes the information easy to grasp for the bot user.
Last but not least, we have built a strong team. We are diverse with respect to age, nationality and background but can work together in a highly cooperative and effective fashion.
The necessities in order to continue the project
The list of credible sources can be expanded, as well as a list of not credible sources. The credibility scores can use semi-supervised machine learning to the problem and include more criteria. The frontend-backend integration can be further developed. The bot’s response is to be tested to give the best trustworthy response.
Log in or sign up for Devpost to join the conversation.